p and p(corr) with missing data for OPLS (Q644)

The information in this article applies to:

SIMCA^®

Summary:

With missing data, p(corr) is not equal to k*p for UV scaled data and OPLS.

More Information:

With UV scaled data and no missing data p(corr)=k*p. With missing data p is calculated with the missing data “filled in” while p(corr) is calculated by excluding the missing cells. This might cause minor differences from a straight line when plotting p vs p(corr) for UV scaled data